A PRESS statistic for working correlation structure selection in generalized estimating equations

Gul Inan; Mahbub A.H.M. Latif; John Preisser

doi:10.1080/02664763.2018.1508560

A PRESS statistic for working correlation structure selection in generalized estimating equations

Gul Inan^*, Mahbub A.H.M. Latif, John Preisser

^*Corresponding author for this work

Department of Mathematics Engineering

Research output: Contribution to journal › Article › peer-review

6 Citations (Scopus)

Abstract

Generalized estimating equations (GEE) is one of the most commonly used methods for regression analysis of longitudinal data, especially with discrete outcomes. The GEE method accounts for the association among the responses of a subject through a working correlation matrix and its correct specification ensures efficient estimation of the regression parameters in the marginal mean regression model. This study proposes a predicted residual sum of squares (PRESS) statistic as a working correlation selection criterion in GEE. A simulation study is designed to assess the performance of the proposed GEE PRESS criterion and to compare its performance with its counterpart criteria in the literature. The results show that the GEE PRESS criterion has better performance than the weighted error sum of squares SC criterion in all cases but is surpassed in performance by the Gaussian pseudo-likelihood criterion. Lastly, the working correlation selection criteria are illustrated with data from the Coronary Artery Risk Development in Young Adults study.

Original language	English
Pages (from-to)	621-637
Number of pages	17
Journal	Journal of Applied Statistics
Volume	46
Issue number	4
DOIs	https://doi.org/10.1080/02664763.2018.1508560
Publication status	Published - 12 Mar 2019

Bibliographical note

Publisher Copyright:
© 2018, © 2018 Informa UK Limited, trading as Taylor & Francis Group.

Funding

This study was initiated while Gul Inan was visiting Department of Biostatistics, University of North Carolina-Chapel Hill, USA. She would like to thank to the Scientific and Technological Research Council of Turkey (TUBITAK) for funding her post-doctoral studies in USA. This study was presented at the 38th Annual Conference of the International Society for Clinical Biostatistics (ISCB) in Vigo, Spain in 2017 through a travel grant program of the ISCB conference. Authors would also like to thank the editor and two anonymous reviewers for their constructive comments which improved the quality of the paper.

Funders	Funder number
International Society for Clinical Biostatistics
TUBITAK
University of North Carolina-Chapel Hill
Türkiye Bilimsel ve Teknolojik Araştırma Kurumu

Keywords

Correlation structure
deletion diagnostics
longitudinal discrete responses
unbalanced longitudinal data
unequally spaced longitudinal data

Access to Document

10.1080/02664763.2018.1508560

Cite this

@article{fbeb9b0372104de8bb01b1c6e3f94a14,

title = "A PRESS statistic for working correlation structure selection in generalized estimating equations",

abstract = "Generalized estimating equations (GEE) is one of the most commonly used methods for regression analysis of longitudinal data, especially with discrete outcomes. The GEE method accounts for the association among the responses of a subject through a working correlation matrix and its correct specification ensures efficient estimation of the regression parameters in the marginal mean regression model. This study proposes a predicted residual sum of squares (PRESS) statistic as a working correlation selection criterion in GEE. A simulation study is designed to assess the performance of the proposed GEE PRESS criterion and to compare its performance with its counterpart criteria in the literature. The results show that the GEE PRESS criterion has better performance than the weighted error sum of squares SC criterion in all cases but is surpassed in performance by the Gaussian pseudo-likelihood criterion. Lastly, the working correlation selection criteria are illustrated with data from the Coronary Artery Risk Development in Young Adults study.",

keywords = "Correlation structure, deletion diagnostics, longitudinal discrete responses, unbalanced longitudinal data, unequally spaced longitudinal data",

author = "Gul Inan and Latif, {Mahbub A.H.M.} and John Preisser",

note = "Publisher Copyright: {\textcopyright} 2018, {\textcopyright} 2018 Informa UK Limited, trading as Taylor & Francis Group.",

year = "2019",

month = mar,

day = "12",

doi = "10.1080/02664763.2018.1508560",

language = "English",

volume = "46",

pages = "621--637",

journal = "Journal of Applied Statistics",

issn = "0266-4763",

publisher = "Routledge",

number = "4",

}

TY - JOUR

T1 - A PRESS statistic for working correlation structure selection in generalized estimating equations

AU - Inan, Gul

AU - Latif, Mahbub A.H.M.

AU - Preisser, John

PY - 2019/3/12

Y1 - 2019/3/12

N2 - Generalized estimating equations (GEE) is one of the most commonly used methods for regression analysis of longitudinal data, especially with discrete outcomes. The GEE method accounts for the association among the responses of a subject through a working correlation matrix and its correct specification ensures efficient estimation of the regression parameters in the marginal mean regression model. This study proposes a predicted residual sum of squares (PRESS) statistic as a working correlation selection criterion in GEE. A simulation study is designed to assess the performance of the proposed GEE PRESS criterion and to compare its performance with its counterpart criteria in the literature. The results show that the GEE PRESS criterion has better performance than the weighted error sum of squares SC criterion in all cases but is surpassed in performance by the Gaussian pseudo-likelihood criterion. Lastly, the working correlation selection criteria are illustrated with data from the Coronary Artery Risk Development in Young Adults study.

AB - Generalized estimating equations (GEE) is one of the most commonly used methods for regression analysis of longitudinal data, especially with discrete outcomes. The GEE method accounts for the association among the responses of a subject through a working correlation matrix and its correct specification ensures efficient estimation of the regression parameters in the marginal mean regression model. This study proposes a predicted residual sum of squares (PRESS) statistic as a working correlation selection criterion in GEE. A simulation study is designed to assess the performance of the proposed GEE PRESS criterion and to compare its performance with its counterpart criteria in the literature. The results show that the GEE PRESS criterion has better performance than the weighted error sum of squares SC criterion in all cases but is surpassed in performance by the Gaussian pseudo-likelihood criterion. Lastly, the working correlation selection criteria are illustrated with data from the Coronary Artery Risk Development in Young Adults study.

KW - Correlation structure

KW - deletion diagnostics

KW - longitudinal discrete responses

KW - unbalanced longitudinal data

KW - unequally spaced longitudinal data

UR - http://www.scopus.com/inward/record.url?scp=85051960661&partnerID=8YFLogxK

U2 - 10.1080/02664763.2018.1508560

DO - 10.1080/02664763.2018.1508560

M3 - Article

AN - SCOPUS:85051960661

SN - 0266-4763

VL - 46

SP - 621

EP - 637

JO - Journal of Applied Statistics

JF - Journal of Applied Statistics

IS - 4

ER -

A PRESS statistic for working correlation structure selection in generalized estimating equations

Abstract

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this